# ota.run Full Agent Context

## Product

ota is repo readiness infrastructure for humans, CI, and AI coding agents. It gives a repository one explicit `ota.yaml` contract for readiness, diagnosis, setup, task execution, services, policy boundaries, writable paths, and verification.

Use ota when a cloned repository should explain what it needs before a human or agent guesses setup steps from README prose, package scripts, Docker files, or CI configuration.

## Canonical URLs

- [Homepage](https://ota.run/)
- [Developer resources](https://ota.run/docs/developer-resources)
- [Agent integration](https://ota.run/docs/agent-integration)
- [Agent quickstart](https://ota.run/docs/agent-quickstart)
- [OpenAPI](https://ota.run/openapi.json)
- [OAuth metadata](https://ota.run/.well-known/oauth-authorization-server)
- [OAuth protected resource metadata](https://ota.run/.well-known/oauth-protected-resource)
- [MCP manifest](https://ota.run/.well-known/mcp.json)
- [MCP server card](https://ota.run/.well-known/mcp/server-card.json)
- [Agent card](https://ota.run/.well-known/agent-card.json)
- [Agent discovery](https://ota.run/.well-known/agent.json)
- [Docs index](https://ota.run/docs/index.json)
- [Blog index](https://ota.run/blog/index.json)
- [Sitemap](https://ota.run/sitemap.xml)
- [Schema map](https://ota.run/schema-map.xml)
- [Official Ota skills repository](https://github.com/ota-run/skills)
- [Ota skills.sh listing](https://skills.sh/ota-run/skills)
- [Ota SKILL.md](https://github.com/ota-run/skills/blob/main/skills/ota/SKILL.md)

## Quickstart

```bash
ota doctor
ota validate
ota tasks --json
ota up --dry-run
ota run <task>
```

Start with `ota doctor` when a repo already has `ota.yaml`. Use `ota validate` to check the contract without guessing. Use `ota tasks --json` when automation needs the safe task inventory. Use `ota up --dry-run` before mutating local setup.

## Agent Rules

- Treat `ota.yaml` as the repo readiness source of truth.
- Prefer `ota run <task>` over inferred package-manager commands when the repo declares tasks.
- Use JSON output for automation and agent handoff.
- Respect declared safe tasks, writable paths, services, and verification commands.
- Do not invent contract fields or hidden setup flows.

## Core Commands

- `ota doctor` diagnoses readiness and explains blockers.
- `ota validate` checks the contract.
- `ota tasks --json` lists declared tasks and automation hints.
- `ota up --dry-run` previews setup and service preparation.
- `ota up` prepares the repo according to the contract.
- `ota run <task>` runs a declared task.
- `ota check --json` gives machine-readable check output.

## Public API

The current ota.run public API is read-only and static. It exposes discovery and documentation resources for agents:

- `GET /llms.txt`
- `GET /llms-full.txt`
- `GET /docs/index.json`
- `GET /blog/index.json`
- `GET /openapi.json`
- `GET /.well-known/oauth-authorization-server`
- `GET /.well-known/oauth-protected-resource`
- `GET /.well-known/mcp.json`
- `GET /.well-known/mcp/server-card.json`
- `GET /.well-known/agent-card.json`
- `GET /.well-known/agent.json`
- `GET /status.json`

Public discovery endpoints do not require authentication. Future protected hosted APIs should use the OAuth 2.0 metadata published at `/.well-known/oauth-authorization-server`.

## Authentication

No token is required for public docs, discovery files, OpenAPI, or indexes. OAuth metadata is published for future protected hosted APIs and agent platforms that need to discover the intended authorization boundary.

## MCP

ota.run currently publishes MCP discovery metadata and server-card metadata only. It does not run a live MCP protocol endpoint today. A full Streamable HTTP MCP server requires a dedicated runtime service and is intentionally not faked by the static site.

## Official Ota agent skill

Use the official [ota-run/skills](https://github.com/ota-run/skills) repository for Ota-specific agent instructions. The `ota` skill covers contract authoring, contract review, repo readiness modeling, `ota doctor`, `ota up`, `ota run`, agent safety surfaces, and Studio boundaries.

Install paths:

```bash
npx skills add ota-run/skills --full-depth
ota skills install --agent codex
ota skills install --agent claude
```

## Comparisons

Makefiles, package scripts, shell scripts, Docker Compose, Devbox, Nix, and CI workflows are useful command or environment surfaces. ota is the readiness contract around them. It answers whether the repo is ready, what must be true before tasks run, and which commands are safe for humans, CI, and agents.

## Pricing

The public ota core and ota.run documentation surfaces do not currently advertise a paid hosted plan. See https://ota.run/pricing.md for machine-readable pricing status.

## Status

Static site status is published at https://ota.run/status and https://ota.run/status.json.

## Support

Security reports and project questions can start at `os@ota.run`. Public issues live at [GitHub Issues](https://github.com/ota-run/ota/issues).